字幕表 動画を再生する
>> All right what I'd like to do today and on Monday is to talk
about NMR spectroscopy and kind of how NMR spectroscopy works.
I'll call it concepts in theory and for me what I want
to do is give my perspective on NMR
which is not a highly mathematical perspective.
In fact, everything I write up here today is going to really be
in terms of numbers is actually going to be simple arithmetic
and most of it is more an embodiment of the idea rather
than a specific calculation that you quote need to do.
So where NMR begins is with the concept that a nucleus
of certain sorts and I'll just write a proton for now,
has a spin to it and when you have a spinning charge it
generates a magnetic dipole.
And if you apply a magnetic field,
we'll call that magnetic field B naught,
then you have two different spin states or more
and you'll see examples of this in the case
of nuclear quadrupoles but let's start with the case
of a proton or a C 13.
You have two spin states that can exist, quanti-spin states.
The spin of the nucleus can either be spin up,
so if it's spin up, in other words in the same direction
as the applied magnetic field then this is going to be lower
in energy so I'll put, by up I mean aligned with B naught
and if it's spin down meaning aligned
against B naught then we're higher in energy and we'll refer
to throughout our discussion.
We'll refer to the lower energy state as the alpha state
and to the higher energy state as the beta state.
Now different types of nuclei have different spin properties.
Rather than trying to start with generalizations
about rules I'll come to those in a moment
because at some point you'll be wondering
in your project well could I study chlorine 35 or something
like that, let's just start with typical nuclei studied.
So if you go for example, to the 400 megahertz NMR spectrometer
in my building in Natural Sciences 1,
you'll find that that instrument can study protons.
I'm going to write a couple of numbers for these.
I'm going to write the atomic number and the mass number.
And it can study C 13 and it can study F 19
and it can study P 31.
And these are common nuclei that are often studied by NMR.
They're easy to study.
What do these nuclei have in common?
They have a one-half indeed and what,
forgetting their spin state what property
on the blackboard do they have in common?
>> Odd numbers of protons and neutrons.
Odd numbers of protons and neutrons
or more specifically we can group them
that their mass number is odd, specifically that the sum
of their protons and neutrons is odd.
So nuclei with an odd mass number have a nuclear spin
and the quantum characterization
of nuclear spin is what's called a spin number
and we'll call the spin number i. It really doesn't matter what
we call it but they call it i and so that number is going
to be one-half and that gives you all the ones up here
but if we want a generalize more nuclei
with an odd mass number will have a spin number of one-half
or three-halves or five-halves, etcetera.
So that's the more general idea.
The ones with one-half are easy
because they have what's called the nuclear dipole.
If you have three-halves or five-halves or one as we'll see
in just a moment you have what's called a nuclear quadrupole
and then those tend to be harder.
So all the ones here of i equals one-half have spin states
so we have the quantum number
and then the two spin states they can have
and so the spin states are plus or minus one-half.
So that's all of these H 1, C 13, oops, F 19 we'll come
to nitrogen in just a second and P 31.
Now a nucleus with a spin number
of three-halves can have spin states of plus or minus one-half
or plus or minus three-halves
and this is what you call a nuclear quadrupole.
Most of the time, many of the times nuclei
with nuclear quadrupoles don't behave as if they're NMR active.
In our next lecture we'll get to the concept of relaxation.
Relaxation basically is how quickly you flip
between the two spin states or in this case,
between the four spin states or three in some cases
and often they flip very quickly
which means you can't study them by NMR.
Relaxation is affected by properties like symmetry as well
and I'll get to that in a moment with another example.
But if I give I an example of a nucleus with a spin state
of three halves, of boron there are two different isotopes.
There are B 10 and B 11 and B 11 has, I think they both do
but B 11 has a spin state of three-halves and if you look
at the NMR spectrum of borohydride
from this one what you see
in the H1 NMR is you see four lines equally spaced
and of equal height due to the hydrogens coupling
with the nuclear quadrupole and it's very unusual
because normally we think about splitting into a doublet
or if you're thinking a triplet or a one to two to one triplet
or quartet or one to three to three to one triplet,
but what's happening here is the hydrogen C boron
and they see either the boron having a spin state
of negative three-halves or negative one-half
or positive one-half or positive three-halves
and so you see the four spin states
and that gives are rise to four lines.
All right but so let's look at some other nuclei
with an odd mass number.
[ Silence ]
So one very important nucleus in biomolecular NMR is N 15.
Nitrogen 15 has a spin number of i equals one
and indeed N 15 is often studied.
Most nitrogens, not N 15.
We talked about this when we talked
about mass spectrometry we said that the natural abundance
of N 15 is 0.38 percent and that's really, really low.
The isotopic abundance of C 13 spin active is one-and a half
percent is 1.1 percent and you know
that carbon NMR is not very sensitive.
You need to have a reasonable sample size, more than you have
for protein typically
and sometimes often collect data for much longer.
Well by the time you're down to .38 percent studying it
at natural abundance is pretty hard so often you do this
with isotopic labeling.
Two dimensional N 15 based techniques are a mainstay
of protein NMR spectroscopy and in general
since most proteins are expressed these days what you do
is you simply grow up your e-coli
with N 15 ammonium chloride and they absorb that and use it
to make up the amino acids
and then you can get a fully N 15 labeled protein
which is very useful.
N 15 is starting to become more important
in some natural product structure determination.
Alkaloids as you may have seen for example
in Neil Gard's [assumed spelling] talks have lots
of nitrogens in them and so being able to figure
out the positions of those nitrogens can be very important.
In the case of something like an alkaloid
or a synthetic project you might not be able to put N 15 in
and NMR spectrometers are becoming more sensitive
and so it becomes not completely nuts to think
about using N 15 techniques in your NMR.
At the end of the course I may talk
about some two the dimensional techniques with N 15
at natural abundance that people are doing just
because I think it's useful but that won't be until the end
of November or December.
Another common, well not common, another nucleus is oxygen,
O 17 remember we said is only low natural abundance.
It's only very low I should say.
It's only.04 percent and oxygen 17 has a spin number
of i is equal to five-halves so that's a nucleus
that can have six spin states, negative five-halves,
negative three-halves, negative one-half, positive one-half,
three-halves, five-halves, etcetera and so it has sort
of doubly damned and so it's not generally studied.
All right so that takes care of our nuclei
with odd mass numbers.
Now the next class I'll talk about is
if you have an even mass number and an even atomic number
so that's easy those are nuclei like C 12,
O 16 and the answer is very simple.
Those have a spin number of i equal zero.
They have no spin and those are NMR inactive.
Since you don't have different spin states you can't have
quanti-transitions between spin states
so there's no way they can be studied by NMR spectroscopy.
So the last class then becomes nuclei with an even mass number
but not atomic number so that would include nuclei
like deuterium, nuclei like N 14.
I guess that would be the common ones we'd encounter
in organic compounds.
These all have a nuclear quadrupole.
Remember a quadrupole is anything
that doesn't have a dipole, i.e. just spin up, spin down,
i.e. i equally one-half so these all have a nuclear quadrupole
and a spin number i equals 1, 2, 3, etcetera so for example,
if you take deuterium you have a spin number i equals 1
and so you have three spin states available to it
and you know the direct manifestation of this that many
of you have seen with your own eyes?
Who's run a C 13 NMR, most of you?
What solvent did you use?
Chloroform, right, the first solvent most of us reach
for because it's pretty cheap as solvents go and pretty good
at dissolving organic chemicals.
It's cheap because it doesn't have
that much deuterium in it, right?
You only have one deuterium for all that weight of chlorine.
You need to deuterium to get the deuterium lock
for NMR spectroscopy and what do you always see
when you run an NMR spectrum in deuterochloroform?
>> A triplet?
>> A triplet and a very interesting sort of triplet
so for CDCl 3 in the C 13 NMR you see a one-to-one-to-one
triplet centered at 77 ppm.
>> These are really jammed together.
>> It's really jammed together.
The separation between the lines is 32 hertz,
in other words the distance
between these two lines is 32 hertz.
The distance between these two lines is 32 hertz.
If you're running your spectrum on a 500 megahertz spectrometer
that means the carbon NMR is running at a 125.7 megahertz.
I'll come back to that in a second
which means 1 ppm is 125 hertz
which means the lines here are separated by about three-tenth
of a ppm and that's on a big roughly 200 ppm scale
so as James said those lines are really close together
and the manifestation is it's a one-to-one-to-one triplet
because to a first order approximation a third
of your deuterons are in spin state negative one.
A third of your deuterons are in spin state zero and a third
of your deuterons are in a spin state of positive one
and we'll see in a moment that they're minuscule,
minuscule differences in the population of the spin states
and that's really, really important.
We'll also see in a moment that that number 32 comes back
when we see something else.
All right most of the time--
so deuterium is kind of special among nuclear quadrupoles
in that most of the time nuclear quadrupoles,
nuclei with quadrupoles undergo rapid relaxation
but deuterium is special.
It's relaxation is slow and I'll just say to put it
in simple terms is effectively like NMR inactive.
So many of the nuclei with nuclear quadrupoles
like chlorine 35 and chlorine 37 how do we know that those have,
all right I will take that back.
We can't know that they have a nuclear dipole
or a nuclear quadrupole but we know they have a spin number
of one-half or three-halves or five-halves or seven-halves.
They happen to have the higher ones so we never see J coupling.
We never see spin-spin coupling to chlorine 35 or chlorine 37,
if we did your carbon spectrum here,
your C 13 NMR spectrum would actually be much more
complicated because you'd be seeing splitting
from the chlorines.
Okay, so nitrogen 15, I'm sorry nitrogen 14 also has a
nuclear quadrupole.
It has a spin number of i equals 1
and so normally you have rapid relaxation, so for example,
if we come to that amide what we were dealing with before
when I asked you about the IR spectrum and if we look
at the NMR spectrum of this amide of course most
of your nitrogen, 99.62 percent
of your nitrogen virtually all is N 14 in here
and we don't see J coupling to this proton
so as I said the fact that we do see J coupling
between the deuterium,
J coupling just means spin-spin coupling to the C 13 is
because deuterium is the odd ball here
and that it often doesn't undergo relaxation
but most nuclei with a nuclear quadrupole don't show nuclear
coupling because we have rapid relaxation.
As I was saying earlier with my example
of borohydride symmetry is the odd ball on
or highly symmetric species end up being odd balls
in that you have slow relaxation
so borohydride B H 4 minus has tetrahedral symmetry
so you see coupling from the boron to the hydrogens a case
that you may see and I saw first by accident one of those cases
where you simply dissolve out the compound in a solution
and you get an NMR spectrum and so this is ammonium chloride
in DMSO, ammonium chloride has NH 4 plus Cl minus
and the ammonium has tetrahedral symmetry
and the first time I happen the accidentally have this
in a sample and took an NMR spectrum as I said in DMSO D6,
I saw an NMR spectrum with 3 peaks that were so far apart
that you barely could tell they went together except the odd
thing was they were all the same height
and this spacing was the same as that and it was what's going on?
Oh, wait that's your nitrogen so this is your J 1 NH.
In other words your one bond coupling between the nitrogen
and the hydrogen and I don't remember what the coupling
constant is but it's big.
>> It was always produced as J no matter?
>> J is, yeah J is the term that we use to refer
to spin-spin coupling.
>> That's not just from proton NMR, right?
>> That's not just proton NMR.
So we would describe this as J 1 CD equals 32 hertz.
And later on when we start to talk
about 2D techniques like HMQC and HMBC.
Terms like J 1 CH, J 2 CH and J 3 CH, in other words one bond,
two bond and three bond carbon hydrogen couplings will become
are very important in structure determination.
All right so when we last left our spinning nuclear dipole he
was spinning in the presence of an applied magnetic field
and I said there were two states, the alpha state
and the beta state and the alpha state was lower in energy
than the beta state so I can make a little diagram,
E and I can show just like you learned in electronic structure
where you learned for example, you have pi orbitals
and pi star orbitals and you have populated electrons
in your two orbital.
Here we can think about populations of nuclei.
It's a little bit different in a sense we're talking
over the entire sample but if we have our applied magnetic field
B naught and we have our alpha state and our beta state,
remember the alpha state is aligned with magnetic field,
we can think about some nuclei being in the alpha state
and some nuclei being in the beta state
and there's an energy gap between these two spin states
and we can think about the energy to flip a nucleus
from the alpha state to the beta state as the energy of a photon,
in other words an energy quantum in the electromagnetic spectrum
and that delta E is going to be H NU.
In other words the energy, the difference, the frequency
of a photon to flip a nucleus from the alpha state
to the beta state is going to be dictated by that difference
in energy such that E equals H NU, delta E equals H NU.
Now what sort of energies are we talking about?
Well we're talking about 500 megahertz for protons
so we're talking about radio frequency,
so let me just give you a calibration here.
So if you think about UV and our delta E so maybe if I think
about UV and I think about a chromophore, maybe I think
about my mercury line at 254 nanometers from my TLC lamp
and I think about a chromophore say containing a benzene ring
or a methoxybenzene ring and maybe I say all right,
if we just take 254 nanometers and I go ahead and plug
into you remember C equals lambda NU
so that's our wavelength and you calculate NU the frequency
and then you calculate E equals H NU and you plug
in the Planck's constant you get the detective to E corresponding
to a photon in the UV is 113 kilocalories per mole.
And then you stop and you think like an organic chemist
and you say okay, wait what's 113 kilocalories per mole?
What's the difference between a pi and a pi star?
It's a little stronger than the strength
of a carbon-carbon single bond, a little stronger
than the strength of a carbon hydrogen bond
in other words the energy difference
in the UV spectrum corresponds to the strength of bonds.
And now if you think about, so this is our UV,
if we think about UV, oh, I guess I wrote UV.
All right if we think about IR and I think
about a typical stretch,
well we've been talking a lot about carbonyls.
Carbonyls absorb at about 1700 wave numbers.
We said that wave numbers was centimeters per wave
which meant your wavelength is 117 hundredth of a centimeter
and that's lambda and then you calculate your frequency out
and it's in the infrared range and then you plug
in to equal delta equals H NU and you find
out that delta E is equal to 4.87 kilocalories per mole.
And you say okay that kind of makes sense.
I know that infrared is lower in energy than UV.
It's lower in energy than visible.
I know that we don't have sufficient energy
to break bonds in the IR.
Indeed all we're doing is kicking them
up a higher vibrational state
and you remember you're energy curves
with your vibrational states
and it takes many jumps before you get to the point
that you're dissociating bonds.
Well if we do the same for NMR
and let's say we take 500 megahertz and we plug in
and again plug in E equals H NU then delta E is equal to 0.0477
but it's not kilocalories per mole.
It's calories per mole.
So the first thing when you see NMR spectroscopy is you're
getting dinged badly because the technique involves very little
energy absorbent.
In other words when you're measuring a UV spectrum it's
very easy for a detector to detect the energy of a photon
and when you're measuring an IR spectrum it's very easy.
And already your detectors have to be much more sensitive
and it's going to get worse from there.
All right so we talked about delta equals H NU,
what's new for a-- that's not a pun,
if it were it would be terrible.
What's new for a nucleus?
NU is dictated by gamma B naught over 2 pi.
Okay, well so far so good.
I said B naught is the applied magnetic field so if you look
at this you'd say well this kind
of makes sense bigger applied magnetic field means bigger
difference in energy, right?
Delta equals H times gamma B naught over 2 pi
so that kind of makes sense.
All right let's just take a look at that.
What does that mean?
There's a linear proportionality,
so if I again plug into this equation I get that,
so in other words if I just go ahead and plug
into this equation I'll come back to gamma in a second.
We find out that if we apply 70, 500 gauss magnetic field
that leads to 300 megahertz for H 1.
If we go to a higher magnetic field that leads
to a higher frequency and it's going to be in a linear fashion
so if I apply 117, 500 gauss magnetic field now we're
at a 500 megahertz NMR spectrometer
and if you make a 300 megahertz NMR spectrometer you have an
electromagnet like this maybe this big,
super-conducting magnet this big where you have a coil of wire
with electricity passing through it, in liquid helium
in the wire is super-conducting so the electricity flows
and flows and flows without any resistance or diminution
and you get a strong magnetic field.
In order to build the technology to get a uniform 117,
500 gauss magnetic field you need a kettle about this big
across and about this high
to house the super-conducting magnet and the liquid helium
and the shims and so forth and finally if you get
to say an 800 megahertz
and of course it's all linear proportionality you're going
to have a 188, 000 gauss magnetic field and that is close
to as big as can currently be made uniform
so now you'll have a magnet that's even bigger and needs
to have its own room in order to house it and flux lines
that go very far out and the limits
on commercial instruments these days are about 900 megahertz
and the thing costs, I guess ours cost about for
to whole thing about 2 and a half million dollars
at a time you're at 900 megahertz it's many,
many millions of dollars and there may be one,
I think one gigahertz out there but we really for now
at least seem to-- what?
>> In France or something.
>> I think so.
We really seem to have just pushed the limits of technology
for how much electricity you can put in a super-conducting coil
without it just ripping itself apart.
All right so the other quantity we have
in this equation is gamma.
Gamma is called the magnetogyric ratio sometimes you'll hear it
referred to as the gyromagnetic ratio.
This is a property of the individual nucleus.
The bigger the gyromagnetic ratio,
the bigger the magnetogyric ratio effectively the bigger the
nuclear spin, the bigger the magnet that the nucleus is.
Protons actually are good.
They have one of the biggest magnetogyric ratio
of any nuclei studied 26, 750 and it's 53.
What am I thinking here?
And so just to put this into context at 117,
500 gauss in other words the relatively large magnet,
so at 117, 500 gauss you have the nuclei flips its spin
at 500 megahertz.
If we look at C 13 we get a gyromagnetic ratio of 6,
728 and that corresponds to absorbing energy at a frequency
of 125.74 megahertz on this 117, 000 gauss magnet.
So one of the implications, remember I said you were dealing
with very small energy differences.
One of the implications is the energy differences are even
smaller for carbon than for proton
so now you're getting doubly damned for carbon
because the national abundance for C 13 is only 1.1 percent
so most of your carbons aren't even C 13.
Indeed with small molecules most
of your molecules don't even contain one C 13 in them.
We saw that in mass spec where you see the C 13 isotopomer peak
and for the small molecules that we were looking
at that peak is small compared to the C 13 isotopomer peak
but you're getting damned again
because its small magnetogyric ratio leads
to smaller energy absorption.
Now the other thing you have
to remember is even though you're recording your C 13 NMR
spectrum on a quote 500 megahertz NMR spectrometer
you're not reporting your carbon NMR on at 500 megahertz,
if you were you'd be that lucky person not in France but maybe
on Mars who has access to a two gigahertz NMR spectrometer
and there ain't no such animal right now.
All right fluorine 19 isn't so bad.
Its magnetogyric ratio is 25, 179 so that corresponds
on the same spectrometer to 470, 000, 470.58 megahertz.
Usually it takes certain types of probe technology.
We'll talk more about that later but certain types
of coil technology to tune to higher frequencies
and certain type of coil technology
to tune to lower frequencies.
So often if you want a really good proton NMR you will use a
special probe where the coil that's tuned for proton is inner
and close to the sample and the coil that's tuned
for other nuclei is bigger and further away from the sample.
That sort of probe won't be as good for carbon 13
because you have the coils further away from the sample,
the coil that's good for C 13.
Conversely, if you find
that Phil Dennison has put a broadband probe
in the spectrometer where the nucleus
at the lower frequency is inside in the coil you may find
that the proton NMR collects is not as good or is not
as sensitive or is not as sharp and well shimmed
because the coil is further out.
Fluorine is interesting because often you can use the same coil
for both fluorine and for proton.
Phosphorus also has a smaller magnetogyric ratio.
It's 10, 840.
Now remember fluorine and phosphorus have all
of their naturally occurring nuclei as F 19 and all
of their naturally occurring nuclei as P 31
so these are not damned
by the low isotopic abundance the way phosphorus is.
Another nucleus that's sometimes studied is deuterium.
Deuterium we talked about the nuclear quadrupole.
You also have your lock coil in there.
The magnetogyric ratio for deuterium is 4,107
so that means your lock frequency
on this spectrometer is at 76.76 megahertz.
[ Silence ]
[ Inaudible student question ]
So the cryoprobe technology is really wonderful.
What they've done
in the cryoprobe technology is they have cooled the probe
and it's either, I guess it's not a super-conducting probe
but what it is is a very low noise probe.
And because the electronics of the probe are cooled
so you don't get much electronic noise
and the result is it's very high sensitivity.
And we were fortunate that had just,
when we bought it they had just developed technology
that had both carbon and proton sensitivity
and basically special coil technology
so that instrument is super good for proton.
It's got a huge, just an incredible signal
to noise ratio, better than even the 800 megahertz spectrometer.
It's also super good for carbon.
I want to come back to these magnetogyric ratios
because you've seen this with your own eyes.
Now we already talked about the coupling, the J 1 CD coupling
in chloroform and remember I said you see this
one-to-one-to-one triplet in the C 13 NMR and the separation
of the lines is 32 hertz.
Well if you've ever looked hard at your chloroform peak
in the proton NMR, so here we have DC coupling,
our H2 coupling but if you ever looked hard
at the chloroform peak
in the proton NMR what you see is something like this.
You see a main peak for your CH
and of course what you're looking at is chloroform
but you'll also see two peaks here
that are the C 13 satellites and those correspond
so this is your C 12 peak and those correspond
to the J coupling to the C 13.
In other words what you're seeing here is a doublet
and the separation of those two lines is 209 hertz
and the mathematical relationship between 209
and 32 is the same as the mathematical relationship,
the ratio between 26,753 and 4107.
In other words it's 6.5,
in other words the magnetogyric ratio is 6.5 times bigger
for a proton than for a deuteron and we see
that directly in the J coupling.
The effect of the magnet that the deuterium has
in splitting the carbon is one-sixth point fifth the effect
that the carbon has in splitting the proton
because coupling is mutual.
All right the last thing I want to talk about I've talked
about how damned we are by energy being low.
I've talked in the case of carbon about isotopic abundance
but now the really damning thing ends
up being the Boltzmann distribution.
That is the population of the spin states.
In the case of a benzene all of your molecules are
in the ground electronic state.
In the case of a ketone all of your carbonyls are
in the ground vibrational state but in the case
of nuclei the energy difference between the alpha
and beta states is so small that both are populated
and if you think back to your P chem
and you calculate the number in the beta state
versus the alpha state that's going to correspond
to the difference in energy, delta E over KT
where K is the Boltzmann constant and then
if we just remember that delta E equals H NU is equal
to H times gamma times B naught over 2 pi and then we say,
okay let's just take at 70,500 gauss, that's our 300 megahertz
and let's plug in N beta divided by N alpha is equal to
and if I plug in that's E to the negative 6.63 times 10
to the negative 34 times the magnetogyric ratio 236753 times
70,500 applied magnetic field over 2 pi divided
by our Planck's constant of 1.38 times 10 to the negative 23
and let's say we're saying at 298 Kelvin.
So I say 298 here and when I work that all
out what I get is this number comes
out to a quotient that's very, very, very close to one .999952.
Four nines and five two corresponds to the ratio
in the beta state over the ratio in the alpha state.
In other words we have 48 more protons out of two million,
so where all of your carbonyls are available to absorb a photon
because remember when you apply a photon it can either kick a
nucleus up from the ground state to the first
from the alpha state to the beta state or down
from the beta state to the alpha state.
So it's only that differential population, only that 48
out of two million that are available to absorb.
If we apply a higher magnetic field it only gets linearly
or almost linearly better.
At 117,500 gauss, remember that's our 500 megahertz then we
only get to an N beta over an N alpha,
in other words a relative population
of point again four nines, .999919.
In other words it only gets a little bit better.
It's only 81 protons out of two million.
So we are damned by the low energies.
We are damned by the low differences in population
and this is why NMR compared
to other spectroscopic techniques is very insensitive
and why it took a long time to develop.
Next time we'll talk about how this NMR spectrometer works,
how we absorb our energies and then how we translate
that into a spectrum and I'll also talk a little bit maybe
about some of the aspects of the spectrum. ------------------------------44553e5ef327--